tqcq/llama.cpp
0
0
Fork 0
mirror of https://github.com/ggml-org/llama.cpp.git synced 2025-06-26 19:55:04 +00:00
Code Issues Packages Projects Releases Wiki Activity

cuda : fix device sync on buffer clear (#14033)

Browse Source
This commit is contained in:
Diego Devesa
2025-06-09 07:36:26 -07:00
committed by GitHub
parent 201b31dc2e
commit 8f47e25f56
1 changed files with 2 additions and 3 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
ggml/src/ggml-cuda/ggml-cuda.cu
View File

@ -615,9 +615,8 @@ static void ggml_backend_cuda_buffer_clear(ggml_backend_buffer_t buffer, uint8_t
ggml_backend_cuda_buffer_context * ctx = (ggml_backend_cuda_buffer_context *)buffer->context;
ggml_cuda_set_device(ctx->device);
CUDA_CHECK(cudaDeviceSynchronize());
CUDA_CHECK(cudaMemset(ctx->dev_ptr, value, buffer->size));
CUDA_CHECK(cudaDeviceSynchronize());
CUDA_CHECK(cudaMemsetAsync(ctx->dev_ptr, value, buffer->size, cudaStreamPerThread));
CUDA_CHECK(cudaStreamSynchronize(cudaStreamPerThread));
}
static const ggml_backend_buffer_i ggml_backend_cuda_buffer_interface = {